Use official to_backend
API for GPU/CPU backend conversion
#767
Loading
to_backend
API for GPU/CPU backend conversion
#767